to measure in terms of social welfare or economic development, they may nonetheless be substantial, accumulative, and inequitably distributed. For example, the increased supply of labor directly depresses the returns to unskilled labor, lowering its relative wage and perhaps increasing the rate of unemployment. If this second component of increased population growth is detrimental to social goals but unlikely to lead to a compensating reduction in fertility, how can the public sector equitably transfer the true social costs of reproduction to the parent who is the ultimate decision-maker?
To reiterate, parents may be provided with the means (birth control) to respond appropriately to the decline in child mortality, but they may not be similarly motivated to compensate for the reduction in adult mortality. Such a change in motivation will require a change in parent demand for surviving children. In order to understand what might generate such a change in the demand of parents for children, this paper began by identifying aspects of the parents' environment that were thought to determine the relative attractiveness of having many versus few children. Some of the more fundamental social and economic changes connected with the postwar development process appear to raise, on balance, the relative costs of rearing children. For tliis reason, one might anticipate that development would slowly foster the adoption of smaller family size goals. Support for the underlying conjecture that fertility is responsive to changing aspects of the parents' environment is drawn from several statistical studies of low income countries.
It is tempting to infer from these exploratory studies that the public sector could effectively hasten the reduction in desired and actual fertility by selective policy measures, such as those aimed at promoting child health, welfare, and education and those assisting women to acquire and employ marketable skills in the paid labor force. But at the moment, understanding of the determinants of family decision-making and its bearing on parent reproductive behavior is too scant to interpret confidently the available evidence. Seen as working hypotheses, however, these inferences might help guide the more extensive micro empirical research and multivariate analysis that is clearly needed in this field. Controversy on the interpretation of the growing supply of micro data on family behavior will probably intensify, not slacken, as different disciplinary perceptions of this broad ecological problem are rigorously formulated and different statistical tools are developed to account for components of observed behavioral, biological, environmental, and cultural change. Though a consensus may not be reached quickly on particulars, there are general grounds for guarded optimism with regard to the micro dynamics of population change.
There are substantial differences in reproductive behavior across regions of low income countries and across individuals in low income communities. The statistical relationships between fertility and the determinants of desired fertility are consistent with the hypothesis that reproductive behavior is respon-